I wanted to share an observation I made when I was discussing with other folks who have attended training, courses, or bootcamps. There is a common feeling that many participants have after the course when they are back at their workplace. They have challenges trying to apply what they have learned in the program. As a fellow trainer, training effectiveness is very important to me and definitely for my clients since they want more bang for their training bucks. It made me very curious why that is the case and I wanted to find out the reason and more.
After much digging, this was what I observed.
Tools
If you look at most of the syllabus and outline, they are very focused on the tools. For instance, most bootcamps will teach linear regression, logistics regression, decision trees, and so on. To make the content more sophisticated they will spend a lot of time to cover the mathematics. For example, how does the decision tree split the group at hand, and how does it come up with the rules for splitting?
You can find throughout the course, that they are focused on teaching all the different machine learning models out there, and I bet my dollars it will also cover neural networks now commonly known as deep learning models, support vector machines, and the mathematics behind them. There is a long list of machine learning models that they have to cover.
And if you do a quick search online, all this content can be found online actually and you might even find a nice video that explains it much better than the instructor!
The analogy I am going to use is, that the course is teaching you how a car engine works! Teaching the tools only covers the “How” part of solving any challenges.
Concepts
Here’s the thing. I believe in teaching concepts rather. Concepts are evergreen (till something drastic changes). Some examples will be best practices, how to select the best models, and how to select the right model performance metrics for the use case.
Concepts are not easy to teach, and it’s something only practitioners with many years of experience might be able to explain well, as they gain more familiarity and establish their own rules and guidelines on solving the challenge, i.e. the heuristics that they establish during the many project experience. You can also see from here, concepts are the most valuable because they cover the “Why” instead, or put in other words, they teach what are the “activation buttons” which then lead to the “how”.
Concept & Tools
To solve the problem of why participants still cannot apply what they learned from the course, the first step is to look at the content of the course. Does the content of the course focus on teaching concepts and tools or just tools only? If the answer to that is yes, the next step is to look at the trainer/instructor of the course. Does the trainer have the necessary experience to cover the concepts thoroughly and explain them well? For this to happen, the trainer must not only have relevant application experience (to know what important concepts to share) but also need training experience as well (to better explain the concepts for participants to understand).
By now, I hope you realize, that to design a good course on data analytics or Artificial Intelligence the content and the trainer’s background are very important. If you decide to attend a short course and expect to be functional in a data analytics or Artificial Intelligence role, that is not the right expectation as there is so much content to cover. And since knowledge of tools can be found online easily, after-course learning is something you have to prepare for, as learning does not stop once the course is finished.
Selecting the right course and trainer, followed by continuous learning is how you can move to being able to apply the tools well, through the concepts learned.
Any thoughts on this you would like to share? Perhaps, something you observed after going through a short course or a boot camp. I will be keen to hear them!
Thank you for reading! Consider supporting my work! You can make a “book” donation and drop me some wisdom! Thanks in advance! :)
Concepts might be simpler to talk about but it can sometimes be pretty hard to integrate to businesses. Some businesses have certain concepts ingrained within their processes (Maybe due to past projects) and injecting a new concept would involve convincing various parties in the company to take it up.
Maybe it'll be easier if higher ups attended such trainings - but even so, the higher ups might sometimes need to take time out to explain the concept to people under their charge - and to make them understand why such concepts make sense...
Maybe ml/ai concepts are a bit more universal and well agreed upon... but Imo, computer science concepts might be sth can be debated about...
Koocept